It is The regression.

If there is no workaround around this problem blame FFmpeg project and its
developers.

After all they little care for users.

On Sat, Feb 21, 2026 at 10:00 AM nolan d via Libav-user <
[email protected]> wrote:

> Hello there,
>
> If we have a filtergraph that is already configured and running, what
> is the best way to change its source?
>
> The details...
>
> 1. In our product, we use a filtergraph that consists of two
> filterchains in series:
>
> front_filterchain -> back_filterchain
>
> 2. Initially, front_filterchain starts with an `anullsrc` and some
> filters. The back_filterchain is always a fixed number of filters
> ending with a `abuffersink`.
>
> 3. The filtergraph is configured and running, and we get 0 samples
> from buffersink as expected.
>
> 4. We then replace front_filterchain with another front_filterchain
> which starts with an `abuffer` and some filters. The back_filterchain
> stays the same.
>
> 5. After calling `avfilter_graph_config()` we use
> `avfilter_config_links()` to configure the new links in
> front_filterchain.
>
> 6. We can then push new frames into the `abuffer` and get processed
> frames out of the `abuffersink`.
>
> 7. This works really well in in our product, which helps hundreds of
> thousands of users exporting their video timeline. We can also mix
> many front_filterchain chains into one back_filterchain and
> dynamically replace them.
>
> front_filterchain1 \
> front_filterchain2 -> back_filterchain
> front_filterchain3 /
>
>
> 8. But, `avfilter_config_links()` was deprecated and removed over the
> course of around last 2 years. I wrote a bit more here:
>
>
> https://lists.ffmpeg.org/archives/list/[email protected]/message/W5LFRAPKMQHEW5RVM2GTAEDNTONUQS7I/
>
> So, what is the best way to go about this. I.e. Hod do we replace
> source filterchains on the fly?
>
> I'm happy to write more about it, if needed. I'm just not sure if this
> is a unique use-case. It feels like a streaming use case that FFmpeg's
> filtergraphs should easily support.
>
> Thanks,
> Nolan
> _______________________________________________
> Libav-user mailing list -- [email protected]
> To unsubscribe send an email to [email protected]
>
_______________________________________________
Libav-user mailing list -- [email protected]
To unsubscribe send an email to [email protected]

Reply via email to